Plan ahead for automation: UNSW academics

By: Andrew Hobbs

Human factors make autonomous trucks too risky in the short term, UNSW researchers say

Plan ahead for automation: UNSW academics
Technology should be used to make it less likely fatigue will occur, rather than addressing its symptoms, the UNSW experts told the Staysafe Committee.


Authorities have not yet considered a range of serious issues with new technologies some hope to rely on in order to make heavy vehicles safer into the future, according to experts from the University of New South Wales’ Transport and Road Safety (TARS) research centre.

The comments were made in the centre’s submission to the NSW parliament’s Joint Standing Committee on Road Safety, also known as the Staysafe Committee, which is holding an inquiry into the use of technology to improve heavy vehicle safety.

Submission authors Professor Ann Williamson and Dr Rena Friswell have been asked to give evidence to the committee in its first hearings, set to start next week, alongside representatives of the Australian Trucking Association, the Livestock, Rural and Bulk Carriers Association and the Australian Logistics Council.

The academics say the practicalities of road transport should temper the enthusiasm that some groups have for technological solutions to safety problems.

In particular, the academics said any assistive technologies which require people to monitor them while they handled the speed and direction of the vehicle should not be used in heavy vehicles, due to human factors.

"It is well-understood that humans are not good at maintaining alertness in monotonous or unstimulating conditions… This is especially a problem in monotonous situations like highway driving," they wrote.

"Currently, technology does not support fully autonomous trucks on the road and it is unlikely fully autonomous truck technology will be available in the foreseeable future."

"Indeed, the range of problems to be resolved has not been fully mapped and new situations that show the weaknesses of various technologies are regularly being identified (e.g., not ‘seeing’ a white truck crossing the road, not predicting kangaroo motion, etc)," they wrote.

Even the launch of partially autonomous vehicles – where drivers might be relieved of driving duty for a period and then asked to resume control later – also create problems, the submission says, because of human factors.

"When drivers have been out-of-the-loop in controlling the vehicle, they do not have an up-to-date understanding of the current task circumstances," they wrote, saying it took the average person between four and six seconds to resume control of a vehicle when they had been otherwise occupied.

"[That is] far too long for a driver to successfully take evasive action," they say.

The impact of automated vehicles on other road users was also important to consider, the authors said, noting that the automated vehicles are much slower than other vehicles at present.

"Large differences in speed between vehicles can increase safety risk. Other drivers become frustrated when the slower vehicle impedes their path so take risky manoeuvres to try to avoid them," they said.

Automated vehicle platoons in particular could make it difficult for other road users to overtake safely, they said, as the length of the platoon would have to be communicated to other road users using a standard protocol – which would require its own public education campaign.

How successful are the fatigue technologies we use today?

Williamson and Friswell also considered the ways fatigue is managed among heavy vehicle drivers in Australia already, saying the problems were mostly due to drivers working for long hours and having little opportunity for rest.

"If heavy truck drivers have active and legal incentives to keep driving rather than to take the time needed for rest and recovery, implementing technology that monitors, detects and warns of fatigue will not be useful," the submission said.

This was particularly important for heavy truck drivers because the size of the vehicle often meant that there were few places where it could stop safely.

"The challenge is to get drivers to respond to their experiences of fatigue early enough which these devices do not do. They activate late in the development of fatigue, often too late for the driver to find an appropriate place to stop."

In particular, the academics said Electronic Work Diaries did not provide all the information that they might – as they did not log the hours a driver worked when they were not behind the wheel – such as when they were refuelling or washing the truck.

"The emphasis of all these technologies is on detecting fatigue when it occurs (treating the symptom) rather than managing work to make it is less likely fatigue will occur (preventing the disease)," the submission said.

"Technological solutions that do not take these contextual issues into account will, at best, only ever deal with secondary indicators of the problem of driver fatigue. They will not help to address the primary causes."


You can also follow our updates by joining our LinkedIn group or liking us on Facebook